Issues in Measuring Representation in the American Legislatures

• In Event: Issues in Representation and State Politics

Abstract

This paper develops estimates of the liberal-conservatism of constituencies, and of general election and primary voters for state legislators and members of Congress. These are based on a variety of data including, for citizens, census data, and church denomination locations in constituencies, along with CCES data. MRP estimates, and variants of more traditionally based sample methods will be compared and contrasted with each other and with estimates derived from a national survey of legislators (25% response rate) which asked the legislators to place their average general election voter and their average primary voter on a 7-point liberalism conservatism scale. Legislators were also asked to place both themselves and their main opponent (if any) in the general election and primary election on the same 7-point scale. In addition to the 7-point liberalism-conservatism items, both the CCES and the legislator survey contain a variety of specific issue items asked identically in both surveys. The analysis will also compare both citizen and legislator placements to existing measures of each (MPR for the former and Nominate and Dime scores for the latter). Thus, the analysis will address the measurement of citizen and legislator ideology and representation in terms of the summary measure of liberalism-conservatism, and in terms of a variety of specific economic and social issues. One of the main issues that will be confronted involves an analysis (Broockman and Skovron) which found that legislators of both parties consistently thought their constituents were more conservative than citizens professed themselves to be on specific issues. If that is true for general electorates and primary electorates, what is the mechanism of effect? Is it, as some might posit, because of the influence of big donors who tend to be conservative? Or are there other explanations? Or are legislators who place voters, rather than their constituency, on the liberalism-conservative scale unbiased rather than systematically biased in those somewhat different placements? Resolving these issues will allow us to examine additional theoretical questions. Under what circumstances, such as uncompetitive general elections, are legislators more likely to represent their primary constituents than their general election constituents. How does constituency heterogeneity affect the representation of constituents? Citizens are generally found to be less ideologically constrained or consistent in the individual issue positions than are legislators. How should this affect our assessment of representation?

Author

• Lynda W. Powell, University of Rochester